Mortality among married older adults in the suburbs of Beirut: estimates from offspring data

• Published in: Bulletin of the World Health Organization 2007-06, Vol.85 (6), p.482-486
• Main Authors: Sibai, AM, Kanaan, MN, Chaaya, Campbell, OMR
• Format: Article
• Language: eng ; por
• Subjects: Health Policy & Services
• ISSN: 0042-9686
• DOI: 10.1590/S0042-96862007000600015

OBJECTIVE: Countries in transition are characterized generally by poor statistical infrastructures and a dearth of vital information. In this study we use offspring data to examine mortality trends in married older men and women through a multipurpose household survey conducted in 2002 in the suburbs of Beirut, Lebanon. The country had been ravaged by war for almost 16 years. METHODS: A random sample of 1520 respondents, with either or both parents surviving their 65th birthday, provided information on 1172 fathers and 1108 mothers. Age- and sex-specific mortality rates per 1000 person-years were estimated. Using log-linear Poisson regression, mortality risk was examined for three birth cohorts: those reaching age 65 before (pre-1975), during (1975-1990) and after (post-1990) hostilities in the country. FINDINGS: A total of 1037 parental deaths were reported, yielding an overall mortality rate of 48.7 per 1000 person-years (51.4 among males and 45.3 among females). Compared to the pre-1975 cohort, older adults reaching age 65 during the war years, 1975-1990, had the highest mortality risk for both males (rate ratio, RR = 1.48, 95% confidence intervals, CI = 1.07-2.04) and females (RR = 1.22, 95% CI = 0.95-1.58). Mortality risk was significantly higher in males than females, a gender differential notably largest in the 1975-1990 cohort. CONCLUSION: This is the first population-based study in Lebanon to quantify patterns of mortality in cohorts of married older adults. The results suggest that the hostilities may have contributed to decreased survival, particularly among males. The approach used in the study presents a viable option for testing in larger surveys and population censuses in countries that lack reliable statistical infrastructures.